OGR data source with driver: ESRI Shapefile 
Source: "/home/joebrew/Documents/cowsquitoes/data/VECTORS/13071_2010_245_MOESM1_ESM/Additional file 1 - African, Euro-ME EO shapefiles", layer: "arabiensis"
with 1 features
It has 1 fields
Integer64 fields read as strings:  ID 

cowsquito: Identifying geographical areas of greatest potential impact for livestock ivermectin implants

Figure 1. Parasite prevalence (a) Raw and (b) percentilized

png 
  2 

Figure 2. Cattle density (a) Raw and (b) percentilized review scale

png 
  2 

Figure 3. Arabiensis density (a) Raw and (b) percentilized

png 
  2 

Figure 4. Combined score just parasite and cattle

png 
  2 

Figure 5. Combined score, all three

png 
  2 

Figure 6. Combined score with first administrative division (a) median score and (b) proportion of surface > median continent score

png 
  2 

OLD STUFF BELOW

Methods

• We use raw raster data on where cattle are from the International Livestock Research Institute (ILRI) and the Food and Agriculture Organization of the United Nations (FAO) and the Université Libre de Bruxelles (ULB-LUBIES).
• We use raw raster data on the Plasmodium falciparum parasite rate in 2-10 year olds in Africa in 2015, made available through the Malaria Atlas Project.
• We use R to process the data, standardize their geographic attributes (extents, projections) and quality attributes (granularity, etc.).
• We use simple percentilization to scale prevalence (0-1) and cattle per square kilometer (0-Inf) into a 0-100 metric.
• We combine our two metrics by simple product So, in the combined score, 0 means no Pf or cattle, and 100 means the maximum amount of Pf and cattle. 0 can also much of one metric, but not of the other (for example, a very malarious region with no cattle gets a 0). Anything between the two represents some combination. This method assumes an equilinear value of both percentilized Pf and cattle (ie, a place with 60th percentile cattle density and 40th percentile Pf is equally as promising as a place with 40th percentile cattle density and 60th percentile Pf).
• Full code in this repo (code.R).

Results

Arabiensis

The below shows the distribution of arabiensis in Africa.

Malaria prevalence on the Plasmodium falciparum parasite rate in 2-10 year-olds.

The below shows the raw data on the Plasmodium falciparum parasite rate in 2-10 year-olds.

The below is identical to the above, but with country borders.

The below is identical to the above, but with administrative borders.

The below shows our (scaled) data on the Plasmodium falciparum parasite rate in 2-10 year-olds.

The below is identical to the above, but with country borders.

The below is identical to the above, but with administrative borders.

Cattle density

The below is identical to the above, but with country borders.

The below is identical to the above, but with administrative level 1 borders.

Combined score

Single score

The below shows our combined score, using the product of the cattle and malaria metrics. Unlike a multi-dimensional score, this visual does not contain as much information (ie, with a value of "40", it's not clear whether this is due to a cattle percentile of 40 and a malaria percentile of 1, or a cattle percentile of 5 and a malaria percentile of 8). That said, a one-dimensional color legend is much easier to interpret visually.

The below is identical to the above, but with country borders.

The below is identical to the above, but with the distribution of arabiensis as well.

The below is identical to the above, but showing a point for each "grid" of the raster surface, and removing the geogrpaphic coordination (ie, scatterplot).

Alternative visualizations

A

B

C

Country aggregation

We aggregate at the polygonal national level so as to rank countries by suitability for intervention. The below maps show the mean, median, maximum, and % of area greater than continental median "single score" (product of Pf and cattle percentile) for each country in Africa.

The below shows a plot of country level values (excluding countries with 0 values throughout).

The below is a table country-level values.

Administrative level 1 aggregation

We aggregate at the polygonal sub-national level so as to rank administrative areas (districts, etc.) by suitability for intervention. The below maps show the mean, median, maximum, and % of area greater than continental median "single score" (product of Pf and cattle percentile) for each administrative unit in Africa.

All values for all sub-national administrative units.

The below shows the top 20 districts in regards to average combined score.

Conclusion

West Africa appears to be the region with the most promise for ivermectin-infused cows. Yummy.

Data sources

Where are the cows?

https://livestock.geo-wiki.org/Application/index.php

Where are the mosquitoes?

(malaria, actually)

https://map.ox.ac.uk/explorer/#/explorer

Contact

Databrew, empowering researchers in academia and industry to explore, understand, and communicate their data through consulting and teaching. info@databrew.cc.